@agentled/cli
Version:
CLI for Agentled — manage workflows, apps, and knowledge from the command line. Zero context-window cost for AI agents.
50 lines (49 loc) • 1.98 kB
JSON
{
"name": "AI with Runtime Tools",
"goal": "AI step that can call web_search + workspace_memory at runtime (aiActionWithTools shape).",
"description": "Starting shape for an AI step that needs to pull fresh web context and recall/store workspace memory. Replace the prompt and response structure for your task.",
"status": "draft",
"context": {
"executionInputConfig": {
"title": "Research + Recall",
"description": "Provide a topic or entity to research.",
"runCTALabel": "Run",
"fields": [
{ "name": "topic", "label": "Topic or entity", "type": "text", "required": true }
]
}
},
"steps": [
{
"id": "start",
"type": "trigger",
"name": "Manual Start",
"pipelineStepStartConditions": { "trigger": { "type": "manual" } },
"next": { "stepId": "analyze" }
},
{
"id": "analyze",
"type": "aiActionWithTools",
"name": "Analyze with Tools",
"tools": [
{ "type": "builtin", "name": "web_search", "builtinType": "web_search" },
{ "type": "builtin", "name": "workspace_memory", "builtinType": "workspace_memory" }
],
"pipelineStepPrompt": {
"template": "Research the topic: {{input.topic}}.\n\nUse `web_search` for fresh external context. Use `workspace_memory` to recall what we already know about {{input.topic}} (call action \"search\" with a relevant query) and to `store` any durable fact you learn (category=fact, confidence 70-100).\n\nReturn a concise structured summary.",
"responseStructure": {
"summary": "string — 3-5 sentences synthesising research + prior memory",
"sources": "array of strings — URLs cited from web_search",
"stored_memories": "array of strings — keys of new memories written (if any)"
}
},
"creditCost": 10,
"next": { "stepId": "done" }
},
{
"id": "done",
"type": "milestone",
"name": "Done"
}
]
}